This one broke for me a few times :/
I actually also think the UI design and feedback mechanisms are a lot worse, so I would recommend that people still use the old one.
As a reminder, here are the main interests of the Fund:
We are particularly interested in small teams and individuals that are trying to get projects off the ground, or that need less money than existing grant-making institutions are likely to give out (i.e. less than ~$100k, but more than $10k). Here are a few examples of project types that we're open to funding an individual or group for (note that this list is not exhaustive):
+ To spend a few months (perhaps during the summer) to research an open problem in AI alignment or AI strategy and produce a few blog posts or videos on their ideas
+ To spend a few months building a web app with the potential to solve an operations bottleneck at x-risk organisations
+To spend a few months up-skilling in a field to prepare for future work (e.g. microeconomics, functional programming, etc).
+ To spend a year testing an idea that has the potential to be built into an org.
This round, due to our much larger budget, I am also interested in applications above $100k in funding. Here are some more examples of things that I would be excited about (though I can’t speak for the other fund members):
- Fund a small research team to work on a particularly important question in the domain of AI Alignment (for example, research similar to AI Impacts work on Moore’s Law and AI timelines, or Wei Dai’s many works in AI Alignment and Decision Theory, or Alex Zhu’s work distilling the Iterated Amplification research agenda).
- Fund a small team to build infrastructure/software for the x-risk research community (like Guesstimate, PredictionBook, LessWrong, Metaculus, or a better version of The Credence Game).
Major Donation: Long Term Future Fund Application Extended 1 Week
2019-02-16T23:30:11.243Z · score: 42 (10 votes)Fixed.
You can sort posts by score by going to the all-posts page and sorting by "top".
I've also thought about this, and would also be interested in seeing examples of this. I would also be interested in someone talking about the legal obstacles to doing this in the U.S. or Europe.
Promoted to curated: I think the linked document is one of the most interesting things to be written in AI Alignment in the last year, and this is the best summary and commentary of it that currently exists. Quality wise, I think everything that I have to say has already been covered by the other commenters, but I overall found reading the linked document, as well as this summary, to be quite helpful in my thinking about AI Alignment, though I also disagree with large parts of it (However, I am not at the research level, and so have a harder time judging how useful it would be for the people who are spending even more time thinking about AI Alignment).
Thanks a lot for writing this summary, and thanks a lot to Eric for all the work he is doing.
My interpretation of Eliezer's position is something like: "If you see an intelligent agent that wasn't specifically optimized away from goal-directedness, then it will be goal-directed". I think I could restate arguments for that, though I remember reading about a bunch of stuff related to this on Arbital, so maybe one of the writeups there gives more background.
Important note: You might lose unsaved comments or posts that are open at the time of the final transition, so I recommend not keeping edits open for super long in the next hour or so (though we can restore them, if you ping us on Intercom).
Yep, seems true to me. I am all in favor of weird exploratory stuff on LW.
Ah, I see. Yeah, I am also not a counterfactual customer who would have ever called them. Before the onset of "standard internet food delivery" I never ordered food, and in the absence of the convenience of that system, would just stop ordering food from any of the restaurants. As such, I am happy to subsidize and give a lot of money to food delivery companies, which seem to actually have created a highly valuable service to me (more so than any individual restaurant ever has for me).
I am somewhat confused by this. I personally much prefer delivery to going out to restaurants, and I don't see why we shouldn't expect a market of delivery food at equally sustainable prices to show up. It seems like on the margin a restaurant is not going to make a loss on delivery food (because if they would, why would they offer it), and I am not a counterfactual customer who would have ever visited their in-person location (except for the very few restaurants that are close-by enough to walk to in 5 minutes).
I would love for most restaurants to drastically reduce the number of their seats and all the other stuff that they can save by prioritizing delivery food, and it seems like I would be defecting on everyone who has similar preferences to me, by sustaining a market of overpriced food that is subsidizing people who go out to restaurants and need all the seating and decorations and stuff.
I think it's already a bit more scalable than what we had before, but I was mostly referring to a future version of the Q&A system, after we added some more related question functionality, and made it so that people notice questions for longer than they stick around on the frontpage, and some way to keep track of major open questions.
My current position is that sometime around now, or maybe in the next two or three months, is the time to experiment with getting more core engagement and generally reaching a broader audience. Q&A in particular is something that I can imagine productively scaling to a larger audience, in a way that actually causes the contributions from the larger audience to result in real intellectual progress.
Ben summarized a lot of my general thoughts on this, but overall I think I agree with a good chunk of your sentiment. I think all the four things you list seem like good things to try, though I am kind of holding off on doing them until I feel more comfortable with the scalability of our intellectual progress machine.
Yeah, I do think having a simple non-anonymous upvoting option is promising. I wonder whether we can make it a simple extension of the strong-upvoting system (maybe have some kind of additional button show up when you strong-upvoted something, that allows you to add your name to it, though I can imagine that getting too complicated UI-wise).
Promoted to curated: You've written a lot of good posts about AI Alignment and related problems, but most of them are a bit too inaccessible to curate, or only really make sense when read together with a lot of other posts. And besides that, I think this post is one of the best and most important ones that I think you've written in the last year, and that I've found to clarify some of my thoughts on this topic, so I think it made sense to curate it.
Thanks a lot for all the work that you are doing, and I hope this post can serve as a good starting point for people who are interested in reading more of your research.
Promoted to curated: I've wanted to curate this for a while, mostly because it introduces a concept in a nice and concise way, while also connecting it to existing concepts and scientific disciplines. I've also found the images pretty useful and good, and I do think we have a lack of images, and see this as a good example of low-effort images increasing the quality of a post by a lot.
Fixed the LaTeX for you. You were in WYSIWYG editor mode, where you type LaTeX by pressing CTRL/CMD and 4 at the same time.
My inner Scott Alexander is screaming: "Have you considered you have ADHD? Have you considered you should take ADHD medicine?"
My inner Scott Alexander does not have a medical degree, and I don't know to what degree I endorse this as advice, but it seemed like the obvious hypothesis on the table. I know of a lot of people who have been helped by ADHD medicine.
I think the situation where you are currently playing soccer, because it's a common sport, and if you would stop doing that you would revert to a sedentary lifestyle, isn't that rare. I've multiple times had the experience that most of my exercise was coming from some kind of hobby, or a constraint of my work-environment (such as being able to walk to work), and that I've reverted to being almost fully sedentary when that hobby or constraint went away, despite attempts at doing plain-exercise or picking up different hobbies.
I do agree, that on the macro-scale, over the course of multiple years, one should be able to find a stable source of exercise that doesn't come with potential head injuries. My comment was more directed at avoiding negative short-term changes in people's lifestyles.
While I agree with this, the consequences of a largely sedentary lifestyle on cognitive performance are also pretty significant, and human motivation is fickle. I am not sure how the calculus turns out, but if you find that the only thing that makes you exercise is playing soccer, then I would probably recommend that you keep playing soccer (while also trying to minimize your use of heading in your soccer strategy).
Edit note: I made some lines headings to make it play nicer with the Table of Contents. Happy to change it back if you are bothered by that.
Usually contracting whitespace helps clean up a lot of formatting glitches that people accidentally introduce, but agree that it's sometimes annoying. I think you should be able to get around it by creating empty paragraphs with whitespace characters in it, or alternatively you could make additional use of dividers.
Yep, sure! I only wanted to be helpful!
Oh, no. Sorry. I didn't intend to make any point. I just thought that having additional examples would be really valuable for this post, and wanted to incentivize people to post them.
Interesting. Maybe also relevant for Kaj's current sequence?
Also: Sorry, this post ended up in our spam-review queue and it took us a few hours to get around to approving it. This means this post just became publicly visible at the top of the frontpage.
Mod note: Made the post an actual link-post and replaced the dots with our dividers. Let me know if you want me to change it back.
Random guess: That's where the post was written.
https://www.youtube.com/watch?v=HcZ48JDamyk&feature=youtu.be
Is at least one.
(Mod note: Made this a normal post instead of a question, since that seemed more in line with the intention. Also, cleaned up some of the formatting.)
I would offer $20 to anyone who can give examples for the three different books in an area, that they actually read, up to a maximum of $60. (e.g. three books in Linear Algebra, or three books in carpeting, or three books about programming).
Oh, the second one seems important. Will fix that ASAP.
The first one makes sense, and should be relatively easy to fix, though it might require resyncing our search database.
Yep, that seems like a bug. It requires some minor changes in architecture to fix, so it might take a bit longer than the smallest bugs.
Promoted to curated: I think this post both communicates a set of important concepts, while also building on past concepts that have been discussed on LessWrong. I think this kind of game-theory is pretty valuable as an intuition pump for a large variety of environments and problems.
Presentation wise, I think the sequences of of agent-behaviors was a bit hard for me to follow, and while I think I was still able to follow the core points without them, finding some way to make them more intuitive or annotate them more would be valuable.
EA Funds: Long-Term Future fund is open to applications until Feb. 7th
2019-01-17T20:27:17.619Z · score: 31 (11 votes)Welcome! :)
I really liked reading your comments on the recent post about Kuhn, and would be interested in hearing more of your thoughts on related topics, And if you ever have any questions or problems with the site, feel free to ping us here or on Intercom!
I know of a lot of people who continued studying and being interested in the forecasting perspective. I think the primary reason why there has been less writing from that is just that LessWrong was dead for a while, and so we've seen less writeups in general. (I also think there were some secondary factors that also contributed, but that the absence of a publishing platform was the biggest)
Welcome!
A lot of people really enjoy Scott Alexander's writing. We have a compilation of the best of his writing here.
A lot of people also really enjoy Harry Potter and the Methods of Rationality, which you can find here.
Another good way of discovering good posts is to go to the all-posts page and sort by "top" which shows you all historical posts sorted by karma. You can find that page with that sorting activated here.
Ah, cool. That makes sense and is quite sad.
I think it's both somewhat of a bug that can be fixed (and I think we've made progress on this in the last year), and just a reality of public communication. I think there is a hierarchy of conversational modes that are more collaborative but require more and more trust among the participants (as Raemon mentioned in his answer), and that conversations usually end up at something close to the lowest level for which there is enough trust among all pairs of participants, which means that any large audience will usually imply a less collaborative communication style.
I think there are some ways around this, and I think the karma system is one of those (by increasing the weight more mutually trusted members have in the conversation, while not fully blocking out newcomers), but it's overall a hard problem and it's not clear to me that it can be solved completely with technical solutions.
Does anyone have examples of multi-stage selection processes, that behave like the one modelled here?
I assign significant probability to Gwern having a list of these already and I just wasn't able to find it, but if there isn't a list, then I will probably just create a new top-level question asking for examples.
Cleaned up some of the formatting, let me know if I broke anything in doing so.
Mod Note: Moved a bunch of comments to the answer section. Ping me if you want it moved back.
Yep, send me a PM with payment details.
I also just went in and appended some random URL parameters to the URL to avoid the duplication filter for now.
Oh, I see, you use the markdown editor.
In that case, I don't think there is currently an easy way to preview it. I will add an issue to our GitHub and see whether I can get around to that soon. Seems good to throw an error when you try to save a post with invalid LaTeX.
Hmm, I think it depends a bit. Do you have a link to a post in which there is LaTeX that displays properly in the editor, and doesn't display properly in the post?
Woops, "implies" is not equal to "is equivalent to". My bad.
Mod edit note: Made this into a question for you. You created it as an ordinary post.
Huh, interesting. I will check that out sometime.
(The NIST randomness beacon doesn't appear to be down to me)